Evaluating Model Precision and Future Data Directions for Bankruptcy Prediction

Table of Links

1. Introduction
2. Literature Review
3. Data Collection
4. Data Analysis
5. Methodology
6. Results
7. Analysis and Interpretation
8. Conclusions and Next Steps, and References

7. Analysis and Interpretation

Due to the complicating factors regarding bankruptcy, it is very unlikely that a model will be generated that will be able to accurately predict all bankruptcies. However, the high precision value that the AutoML model generated gives us hope that an accurate model can be created.

8. Conclusions and Next Steps

We recommend more bankruptcy data be collected, particularly more recent data. Furthermore, AutoML may be a more cost -saving method of modeling. It allows users to implement machine learning processes with greater ease than running different models with associated parameters manually.

The AutoML process is fairly simple: upload the CSV data or connect to a BigQuery table, pick a target column for prediction, select a metric to optimize, and select an amount of training hours to run the model .

The 2020 -2021 COVID pandemic had an unseen impact on the fallen -angel bond market. Credit rating agencies have downgraded many companies that were deemed at risk of being unable to repay investors when their bonds mature. By the end of 2020, the fallen -angel universe more than doubled compared to pre -pandemic. Many of these bonds have outperformed the overall high - yield market and investment -grade bonds in the last decade. Buying fallen -angel bonds in the clients ’ portfolios can yield a significant return when these companies improve their finances and upgrade back to investment grade. When this occurs, the prices will spike when these bonds are again eligible for investment in the investment group portfolios. In the first half of 2022, about $69 billion of high -yield bonds were upgraded to investment grade (Veraa, 2022).

This information above represents a great opportunity to take the research to the next level. By collecting the data on these fallen angel bonds, we would be able to gain a higher set of data regarding companies that go bankrupt than ever before. This should be an area to direct the next data science team toward regarding data collection.

References

Ahn, B.A., S.S. Cho, and C.Y. Kim, 2000. “The integrated methodology of rough set theory and artificial neural network for business failure prediction.” Expert Systems with Applications, 18 (2000), pp. 65-74, https://doi.org/10.1016/S0957-4174(99)00053-6

Altman, Edward I. (1968). Financial ratios, discriminant analysis and the prediction corporate bankruptcy. Journal of Finance, 23(4), 589–609.

Altman, Edward I. and Gabriele Sabato, 2008. “Modelling credit risk for SMEs: evidence from the U.S. market.” Abacus, 43 (43) (2007), pp. 332-357. https://doi.org/10.1111/j.1467-6281.2007.00234.x.

Brownlee, Jason, 2020. “SMOTE for Imbalanced Classification with Python.” Machine Learning Mastery. https://machinelearningmastery.com/smote-oversampling-for-imbalanced-classification/.

DataTechNotes, 2021. “SelectKBest Feature Selection Example in Python.” https://www.datatechnotes.com/2021/02/seleckbest-feature-selection-example-in-python.html.

Devatha, Vikram, 2019. “Predicting bankruptcy using Machine Learning | by Vikram Devatha.” Towards Data Science. https://towardsdatascience.com/predicting-bankruptcy-f4611afe8d2c

Dubey, Akash, 2018. “Feature Selection Using Random forest | by Akash Dubey.” Towards Data Science. https://towardsdatascience.com/feature-selection-using-random-forest-26d7b747597f.

Edmister, Robert O., 1972. “An empirical test of financial ratio analysis for small business failure prediction.” Journal of Financial and Quantitative Analysis, 7 (2) (1972), pp. 1477-1493, https://doi.org/10.2307/2329929.

Fletcher, D. and E. Goss, 1993. “Application forecasting with neural networks an application using bankruptcy data.” Information and Management, 24 (1993), pp. 159-167

Gang, Kou, Yong Xu, Yi Peng, Feng Shen, Yang Chen, Kun Chang, and Shaomin Kou. 2021. “Bankruptcy prediction for SMEs using transactional data and two-stage multiobjective feature selection.” Decision Support Systems, Volume 140, ISSN 0167-9236. https://doi.org/10.1016/j.dss.2020.113429.

Grunert, Jens, Lars Norden, and Martin Weber, 2005. “The role of non-financial factors in internal credit ratings.” J. Bank. Finance, 29 (2) (2005), pp. 509-531. https://doi.org/10.1016/j.jbankfin.2004.05.017.

Liang, Deron, Chia-Chi Lu, Chih-Fong Tsai, and Guan-An Shih, 2016. “Financial Ratios and Corporate Governance Indicators in Bankruptcy Prediction A Comprehensive Study.” European Journal of Operational Research 252 (2016) 561- 572. https://www.sciencedirect.com/science/article/pii/S0377221716000412

Menon, Kartik, 2022. “Feature Selection In Machine Learning.” Simplilearn. https://www.simplilearn.com/tutorials/machine-learning-tutorial/feature-selection-in-machine-learning.

Scikit learn, n.d. “1.13. Feature selection - scikit-learn 1.1.2 documentation.” Accessed October 26, 2022. https://scikitlearn.org/stable/modules/feature_selection.html#univariate-feature-selection.

Scikit-learn: Machine Learning in Python, Pedregosa et al., JMLR 12, pp. 2825-2830, 2011.

Veraa, Karen, 2022. “Why bond investors may find sanctuary in ‘fallen angels.’” iShares. https://www.ishares.com/us/insights/bonds-fallen-angels-investment.

Wu, Ye. 2022. “7 Techniques to Handle Imbalanced Data.” KDnuggets. <https://www.kdnuggets.com/2017/06/7- techniques-handle-imbalanced-data.html>.